1 1 00:00:00,480 --> 00:00:02,400 Network performance issues. 2 2 00:00:02,400 --> 00:00:04,050 In this video, we're going to discuss 3 3 00:00:04,050 --> 00:00:06,090 how to troubleshoot network performance issues 4 4 00:00:06,090 --> 00:00:08,180 that you may experience out in the field. 5 5 00:00:08,180 --> 00:00:09,860 Now the five most common causes 6 6 00:00:09,860 --> 00:00:11,320 of network performance issues 7 7 00:00:11,320 --> 00:00:13,650 are high CPU usage in a network device, 8 8 00:00:13,650 --> 00:00:15,610 high bandwidth usage in a network, 9 9 00:00:15,610 --> 00:00:17,130 poor physical connectivity, 10 10 00:00:17,130 --> 00:00:20,500 malfunctioning of your network or DNS problems. 11 11 00:00:20,500 --> 00:00:23,090 First, it's important to realize that every network device 12 12 00:00:23,090 --> 00:00:25,530 such as your routers, switches and firewalls 13 13 00:00:25,530 --> 00:00:27,380 are at their core, a computer. 14 14 00:00:27,380 --> 00:00:30,350 That includes a CPU or central processing unit. 15 15 00:00:30,350 --> 00:00:31,900 If that device begins to experience 16 16 00:00:31,900 --> 00:00:33,930 a high CPU utilization rate, 17 17 00:00:33,930 --> 00:00:36,760 that device is going to start slowing down and in turn, 18 18 00:00:36,760 --> 00:00:39,470 the network that is being processed for all those packets 19 19 00:00:39,470 --> 00:00:41,310 is also going to slow down. 20 20 00:00:41,310 --> 00:00:43,870 So whenever a CPU is overutilized, 21 21 00:00:43,870 --> 00:00:47,340 the latency, jitter and packet loss will start to increase 22 22 00:00:47,340 --> 00:00:49,630 and this can cause the entire network performance 23 23 00:00:49,630 --> 00:00:51,210 to deteriorate on you. 24 24 00:00:51,210 --> 00:00:53,610 So to solve this issue, you need to either 25 25 00:00:53,610 --> 00:00:56,030 upgrade your network devices to more powerful ones 26 26 00:00:56,030 --> 00:00:58,730 such as enterprise class routers and firewalls 27 27 00:00:58,730 --> 00:01:01,250 or you need to simplify the processing load 28 28 00:01:01,250 --> 00:01:03,190 that you're putting on those devices. 29 29 00:01:03,190 --> 00:01:05,710 For example, I once saw an enterprise network 30 30 00:01:05,710 --> 00:01:07,000 where their border firewall 31 31 00:01:07,000 --> 00:01:09,240 was slowing down the entire network. 32 32 00:01:09,240 --> 00:01:10,830 Now when we looked at it closer, 33 33 00:01:10,830 --> 00:01:14,390 we saw their access list had over 3,000 entries in it. 34 34 00:01:14,390 --> 00:01:16,720 That meant that every time a packet was received, 35 35 00:01:16,720 --> 00:01:19,520 it was checked against these 3,000 individual entries 36 36 00:01:19,520 --> 00:01:23,010 to find its match or to eventually block that packet. 37 37 00:01:23,010 --> 00:01:25,450 Now this consumes a lot of processing time 38 38 00:01:25,450 --> 00:01:28,100 and leads to high CPU utilization rates. 39 39 00:01:28,100 --> 00:01:29,900 Now we conducted an ACO review 40 40 00:01:29,900 --> 00:01:31,680 and we were able to simplify their rule set 41 41 00:01:31,680 --> 00:01:33,600 down to about 300 rules. 42 42 00:01:33,600 --> 00:01:36,430 This drastically reduced the CPU utilization rate they had 43 43 00:01:36,430 --> 00:01:37,990 and sped up the network. 44 44 00:01:37,990 --> 00:01:39,990 Second, high bandwidth utilization 45 45 00:01:39,990 --> 00:01:43,000 is another cause of network performance deterioration. 46 46 00:01:43,000 --> 00:01:45,250 Now when bandwidth utilization is high, 47 47 00:01:45,250 --> 00:01:46,700 network communications have to wait 48 48 00:01:46,700 --> 00:01:48,590 to enter or leave that network. 49 49 00:01:48,590 --> 00:01:50,190 This can cause buffers to fill up 50 50 00:01:50,190 --> 00:01:53,010 and in the worst cases, packets can be dropped. 51 51 00:01:53,010 --> 00:01:54,750 Now when those packets are dropped, 52 52 00:01:54,750 --> 00:01:57,750 they're going to be retransmitted if they were sent using TCP. 53 53 00:01:57,750 --> 00:01:58,920 And this leads again, 54 54 00:01:58,920 --> 00:02:01,020 to even higher bandwidth utilization rates 55 55 00:02:01,020 --> 00:02:02,920 because we have retransmissions. 56 56 00:02:02,920 --> 00:02:04,010 To solve this issue, 57 57 00:02:04,010 --> 00:02:05,840 you can either increase your bandwidth size 58 58 00:02:05,840 --> 00:02:07,960 by paying more to your internet service provider 59 59 00:02:07,960 --> 00:02:10,470 or you can do a network flow analysis 60 60 00:02:10,470 --> 00:02:12,560 and determine what sites and traffic types 61 61 00:02:12,560 --> 00:02:14,600 are being used by all of your clients. 62 62 00:02:14,600 --> 00:02:17,350 For example, if 90% of your traffic is being wasted 63 63 00:02:17,350 --> 00:02:19,340 by people scrolling Facebook each day, 64 64 00:02:19,340 --> 00:02:21,660 you may need to change your acceptable use policy 65 65 00:02:21,660 --> 00:02:22,800 or reduce or eliminate 66 66 00:02:22,800 --> 00:02:25,310 when people are allowed to use Facebook on your network. 67 67 00:02:25,310 --> 00:02:28,400 Literally, I had one network where 90% of our traffic 68 68 00:02:28,400 --> 00:02:30,880 was either going to or coming from Facebook 69 69 00:02:30,880 --> 00:02:32,450 and our end users kept complaining 70 70 00:02:32,450 --> 00:02:34,760 because they couldn't do any real work on their job 71 71 00:02:34,760 --> 00:02:36,710 because our internet was so slow. 72 72 00:02:36,710 --> 00:02:38,970 And this is why conducting a NetFlow analysis 73 73 00:02:38,970 --> 00:02:40,090 can really help you understand 74 74 00:02:40,090 --> 00:02:42,200 what your users are really doing on your network 75 75 00:02:42,200 --> 00:02:44,460 and is it something you want to allow. 76 76 00:02:44,460 --> 00:02:46,790 Third, you may have network performance issues 77 77 00:02:46,790 --> 00:02:48,750 if you have a poor physical connection. 78 78 00:02:48,750 --> 00:02:51,030 To solve this, you're going to need to check your cables 79 79 00:02:51,030 --> 00:02:52,850 and test them one by one. 80 80 00:02:52,850 --> 00:02:54,310 Now if you suspect it's an issue 81 81 00:02:54,310 --> 00:02:56,570 with the internet search provider's portion of the network, 82 82 00:02:56,570 --> 00:02:57,970 you need to connect a test client 83 83 00:02:57,970 --> 00:02:59,670 directly to the demarcation point 84 84 00:02:59,670 --> 00:03:01,610 and test the connection from there. 85 85 00:03:01,610 --> 00:03:03,080 If that connection is poor, 86 86 00:03:03,080 --> 00:03:05,330 then you know it's your ISP's problem to solve. 87 87 00:03:05,330 --> 00:03:07,620 If it isn't, that means it's something in your network 88 88 00:03:07,620 --> 00:03:10,280 that's the problem and you need to start testing every cable 89 89 00:03:10,280 --> 00:03:11,420 from the from the demarc point 90 90 00:03:11,420 --> 00:03:13,230 all the way back to your border router 91 91 00:03:13,230 --> 00:03:15,200 and then to your switch and then to your client 92 92 00:03:15,200 --> 00:03:18,050 until you find where that broken or damaged cable might be 93 93 00:03:18,050 --> 00:03:20,580 that's causing the slowdown in your network connection. 94 94 00:03:20,580 --> 00:03:23,260 Remember, a damaged cable may still operate, 95 95 00:03:23,260 --> 00:03:24,910 but it's going to cause additional errors 96 96 00:03:24,910 --> 00:03:26,270 and these errors require data 97 97 00:03:26,270 --> 00:03:28,310 to be retransmitted over and over again 98 98 00:03:28,310 --> 00:03:30,930 and this slows down your overall network performance. 99 99 00:03:30,930 --> 00:03:33,440 To test your cables, always use a cable tester 100 100 00:03:33,440 --> 00:03:34,820 for twisted pair connections 101 101 00:03:34,820 --> 00:03:38,070 or a fiber light meter for fiber optic connections. 102 102 00:03:38,070 --> 00:03:40,750 Fourth, we have malfunctioning of our network. 103 103 00:03:40,750 --> 00:03:43,020 Now if you have misconfigurations of your devices 104 104 00:03:43,020 --> 00:03:44,260 or hardware failures, 105 105 00:03:44,260 --> 00:03:47,050 your network performance is obviously going to suffer. 106 106 00:03:47,050 --> 00:03:48,780 Sometimes this can even be caused 107 107 00:03:48,780 --> 00:03:51,400 by using old or outdated network operating systems. 108 108 00:03:51,400 --> 00:03:52,750 Again, this is an area 109 109 00:03:52,750 --> 00:03:54,500 where your seven step troubleshooting method 110 110 00:03:54,500 --> 00:03:56,700 is going to help you identify the exact location 111 111 00:03:56,700 --> 00:03:58,750 of the device that's having these issues 112 112 00:03:58,750 --> 00:04:01,790 and then you can focus on whether it's a configuration issue 113 113 00:04:01,790 --> 00:04:03,260 or a hardware failure issue 114 114 00:04:03,260 --> 00:04:05,530 that's causing those network performance issues. 115 115 00:04:05,530 --> 00:04:08,920 Finally, fifth, we have DNS problems in our network. 116 116 00:04:08,920 --> 00:04:11,170 Now we already talked specifically about DNS issues 117 117 00:04:11,170 --> 00:04:12,310 in a separate lesson, 118 118 00:04:12,310 --> 00:04:14,810 but DNS issues are a serious cause 119 119 00:04:14,810 --> 00:04:16,940 of a lot of our network performance issues. 120 120 00:04:16,940 --> 00:04:19,650 Remember, if you have a high DNS latency, 121 121 00:04:19,650 --> 00:04:22,020 this is going to slow down the overall user experience 122 122 00:04:22,020 --> 00:04:24,050 because when that user requests a website, 123 123 00:04:24,050 --> 00:04:26,370 they first have to go and resolve that domain name 124 124 00:04:26,370 --> 00:04:28,000 all the way out to that DNS server 125 125 00:04:28,000 --> 00:04:30,393 and get the IP address over DNS.